The attachment of spindle microtubules to kinetochores is crucial for accurate segregation of chromosomes to daughter cells during mitosis. While a growing number of proteins involving this step are being identified, its molecular mechanisms are still not clear. We found that atypical protein kinase (aPKC) is localized at the mitotic spindle during mitosis and plays a role in stable kinetochore-microtubule attachment. Striking staining for aPKC was observed at the mitotic spindle and spindle poles in cells at prometaphase and metaphase. aPKC molecules at these stages were phosphorylated at Thr-410, as detected by a phosphospecific antibody. aPKC was also detected at the spindle midzone and the midbody during anaphase and telophase, respectively, and aPKC at these stages was no longer phosphorylated at Thr-410. The polarity determinants Par3 and Par6, which are known to associate with aPKC, were colocalized with aPKC at prometaphase and metaphase, but not at anaphase and telophase. Knockdown of aPKC by RNA interference affected normal chromosome alignment leading to generation of cells with aberrant nuclei. A specific aPKC inhibitor strongly blocked the formation of cold-sensitive stable kinetochore microtubules, and thus prevented microtubule-kinetochore attachment. Treatment of cells with the aPKC inhibitor also dislocated the minus-end directed motor protein dynein from kinetochores, but not the mitotic checkpoint proteins Mad2 and CENP-E. Prolonged exposure to the aPKC inhibitor eventually resulted in cell death. These results suggest a critical role of aPKC in spindle microtubule-kinetochore attachment and subsequent chromosomal separation. Cell polarity regulates diverse biological events such as localization of embryonic determinants and establishment of tissue and organ architecture. Epithelial cell polarity is regulated by the polarity complex Par6/Par3/atypical protein kinase C (aPKC). We previously found that the nucleotide exchange factor ECT2 associates with this polarity complex and regulates aPKC activity, but the role of ECT2 in cell polarity is still unclear. In this fiscal year we found that expression of a dominant negative (ECT2-N2) or constitutively active (ECT2-&#916;N5) form of ECT2 inhibits normal cyst formation of MDCK cells in 3-dimentional collagen gels. Central lumens were not observed in cysts formed by cells expressing either ECT2-&#916;N5 or ECT2-N2. Apical localization of ZO-1 and basolateral localization of &#946;-catenin were no longer observed in these cells. Interestingly, cells expressing ECT2-N2 did form normal cysts when cultured in the basement membrane matrix Matrigel instead of collagen gels. Addition of a major Matrigel component, laminin, partially rescued the normal cyst formation inhibited by ECT2-N2 in 3-dimentional collagen gels. Thus, signaling through laminin might override the defects of signaling through collagen and ECT2. Whereas ECT2-N2 inhibited the lumen formation of MDCK cysts, caspase-3, which is reportedly involved in lumen formation through apoptosis, was activated at various locations of cells in the cysts. It is likely that perturbation of ECT2 signaling inhibits the establishment of epithelial cell polarity leading to the inhibition of selected elimination of cells at the center of cysts. Thus, ECT2 appears to play a critical role in epithelial cell polarity.